Logloss increasing with num_rounds in distributed + tree_method=exact


We have observed a behaviour that when

  • tree_method set to exact
  • distributed mode
  • trainig with same parameters only increasing the number of trees

The logloss metric is worsening (rising) with more trees built.

Is this a known bug of exact+distributed or should I create an issue?


Please create an issue with a reproducible script.


I have created an issue here https://github.com/dmlc/xgboost/issues/5304